What will a future without secrets look like? | Alessandro Acquisti

翻译人员: xuan wang 校对人员: Jing Peng

I would like to tell you a story

我想跟大家分享一个

connecting the notorious privacy incident

将亚当和夏娃的

involving Adam and Eve,

臭名昭著的隐私事件

and the remarkable shift in the boundaries

和过去十年发生的

between public and private which has occurred

公共和隐私领域里的显著变迁

in the past 10 years.

相结合的故事。

You know the incident.

大家都知道这个事件吧。

Adam and Eve one day in the Garden of Eden

在伊甸园里有一天亚当和夏娃

realize they are naked.

意识到了他们是赤裸的。

They freak out.

他们吓坏了。

And the rest is history.

然后剩下的就是历史了。

Nowadays, Adam and Eve

换作现在的话，亚当和夏娃

would probably act differently.

可能会有不同的举动。

[@Adam Last nite was a blast! loved dat apple LOL]

（推特）［＠亚当，昨天太销魂了！我好爱那个苹果啊。］

[@Eve yep.. babe, know what happened to my pants tho?]

［＠夏娃，是啊，宝贝儿，知道我裤子变成什么样了吗？］

We do reveal so much more information

我们确实比以往任何时候都开放,

about ourselves online than ever before,

把大量关于自己的信息放在网上传播。

and so much information about us

而且这么多有关我们的信息

is being collected by organizations.

正在被各种机构收集起来。

Now there is much to gain and benefit

当今,通过对这些大量

from this massive analysis of personal information,

个人信息的研究,

or big data,

我们从中受益非浅;

but there are also complex tradeoffs that come

但是在放弃我们的隐私的同时

from giving away our privacy.

也要付出很多的代价。

And my story is about these tradeoffs.

而我的故事就是关于这些代价的。

We start with an observation which, in my mind,

让我们首先看看一个我认为

has become clearer and clearer in the past few years,

在过去几年已经变得越来越清晰的现象,

that any personal information

那就是任何个人信息

can become sensitive information.

都可能变成敏感信息。

Back in the year 2000, about 100 billion photos

在2000年的时候，全球大约拍摄了1000亿

were shot worldwide,

的照片，

but only a minuscule proportion of them

但是只有非常微不足道的一部分

were actually uploaded online.

被放在了网上。

In 2010, only on Facebook, in a single month,

到了2010年，仅仅在脸书上，一个月

2.5 billion photos were uploaded,

就上传了25亿张照片，

most of them identified.

大部分都是可确认的。

In the same span of time,

同时，

computers' ability to recognize people in photos

计算机在照片中识别面孔的能力

improved by three orders of magnitude.

提高了三个数量级。

What happens when you combine

当你把这些技术结合起来后

these technologies together:

会发生什么呢？

increasing availability of facial data;

不断增加的脸部信息可用性;

improving facial recognizing ability by computers;

不断增强的计算机面部识别能力；

but also cloud computing,

同时还有云计算，

which gives anyone in this theater

让在座的任何人

the kind of computational power

都拥有了

which a few years ago was only the domain

几年前只有情报机构才有的

of three-letter agencies;

计算能力；

and ubiquitous computing,

同时还有普适计算，

which allows my phone, which is not a supercomputer,

让我的手机，

to connect to the Internet

和互联网相连接

and do there hundreds of thousands

然后可以在几秒内进行

of face metrics in a few seconds?

成百上千的面部数据测算。

Well, we conjecture that the result

我们预测

of this combination of technologies

这些技术的结合体

will be a radical change in our very notions

会对我们所谓的隐私和匿名

of privacy and anonymity.

产生非常巨大的影响。

To test that, we did an experiment

为了证明这个想法，我们在

on Carnegie Mellon University campus.

卡内基·梅隆大学校园里做了一个测试。

We asked students who were walking by

我们让过路的学生们

to participate in a study,

参与一项研究，

and we took a shot with a webcam,

我们用摄像头给他们照了相，

and we asked them to fill out a survey on a laptop.

然后我们让他们在电脑上填写一张调查问卷。

While they were filling out the survey,

在此同时,

we uploaded their shot to a cloud-computing cluster,

我们把他们的照片上传到一个云计算节点上，

and we started using a facial recognizer

然后我们开始用一个面部识别程序来

to match that shot to a database

把那张照片和一个有

of some hundreds of thousands of images

成百上千张照片的数据库相比较对照

which we had downloaded from Facebook profiles.

这些照片都是我们从脸书上下载下来的。

By the time the subject reached the last page

当被研究对象做到问卷的最后一页时,

on the survey, the page had been dynamically updated

那页已经自动显示我们找到的

with the 10 best matching photos

10张由识别程序找到的

which the recognizer had found,

最相似的图片，

and we asked the subjects to indicate

然后我们让被研究对象确认

whether he or she found themselves in the photo.

那些照片到底是不是自己。

Do you see the subject?

大家看到被研究对象了吗？

Well, the computer did, and in fact did so

电脑做到了，实际上它的准确率是

for one out of three subjects.

三分之一。

So essentially, we can start from an anonymous face,

基本上，我们可以从一张匿名的面孔开始，

offline or online, and we can use facial recognition

线下或线上，然后我们可以用脸部识别技术

to give a name to that anonymous face

找到那个人。

thanks to social media data.

这多亏了社交媒体的数据。

But a few years back, we did something else.

但是几年前，我们做了些其他事情。

We started from social media data,

我们从社交媒体数据出发，

we combined it statistically with data

然后我们把它和美国政府的

from U.S. government social security,

社会安全机构里的数据相对照，

and we ended up predicting social security numbers,

我们最终可以预测一个人的社会保险号码，

which in the United States

这个号码在美国

are extremely sensitive information.

是极其敏感的信息。

Do you see where I'm going with this?

大家明白我的意思了吗？

So if you combine the two studies together,

如果你把这两个研究相结合，

then the question becomes,

问题就来了，

can you start from a face and,

你可不可以从一张面孔出发，

using facial recognition, find a name

然后通过面部识别找到这个人

and publicly available information

和有关此人的

about that name and that person,

各种公共信息，

and from that publicly available information

从这些公共信息里，

infer non-publicly available information,

可以推断出未公开的信息，

much more sensitive ones

即那些关于此人

which you link back to the face?

更敏感的信息呢？

And the answer is, yes, we can, and we did.

答案是，可以的，我们也做到了。

Of course, the accuracy keeps getting worse.

当然，准确率也变糟了。

[27% of subjects' first 5 SSN digits identified (with 4 attempts)]

［27％的调查对象的社会保障号头5个数字 可以通过4次尝试得到］

But in fact, we even decided to develop an iPhone app

但实际上，我们甚至决定开发一个苹果应用,

which uses the phone's internal camera

这个应用使用手机内置的相机给

to take a shot of a subject

研究对象拍照

and then upload it to a cloud

然后把照片上传到云端

and then do what I just described to you in real time:

然后实时地进行我刚才描述的计算：

looking for a match, finding public information,

寻找匹配，公共信息，

trying to infer sensitive information,

尝试推测敏感信息，

and then sending back to the phone

然后把这些信息传送回手机

so that it is overlaid on the face of the subject,

然后把这些信息列到研究对象的图像旁边,

an example of augmented reality,

这是个夸张现实的例子，

probably a creepy example of augmented reality.

大概也是一个令人毛骨悚然的现实。

In fact, we didn't develop the app to make it available,

实际上，我们没有开发这个应用，

just as a proof of concept.

这只是一个概念验证。

In fact, take these technologies

事实是，让我们把这些技术推进到

and push them to their logical extreme.

逻辑的极限。

Imagine a future in which strangers around you

设想一下未来你周围的陌生人

will look at you through their Google Glasses

可以通过他们的谷歌眼镜

or, one day, their contact lenses,

或者，他们的隐形眼镜，

and use seven or eight data points about you

并通过你身上的7、8个数据点

to infer anything else

就可以推测出

which may be known about you.

任何与你有关的信息。

What will this future without secrets look like?

这个没有任何秘密的未来会是怎样的？

And should we care?

而我们该不该关心这个问题？

We may like to believe

我们可能会倾向于相信

that the future with so much wealth of data

这个有这么丰富的数据的未来

would be a future with no more biases,

会是一个不再有偏见的未来,

but in fact, having so much information

但实际上，拥有这么多的信息

doesn't mean that we will make decisions

并不意味着我们就会做出

which are more objective.

更理性的选择。

In another experiment, we presented to our subjects

在另一个试验里，我们给研究对象

information about a potential job candidate.

关于一个工作应征者的信息。

We included in this information some references

上传信息里同时也包括了一些

to some funny, absolutely legal,

有趣并且绝对合法，

but perhaps slightly embarrassing information

但毕竟有些

that the subject had posted online.

尴尬的内容。

Now interestingly, among our subjects,

有趣的是，

some had posted comparable information,

一部分研究对象发布了类似的信息，

and some had not.

有些没有。

Which group do you think

大家认为哪个组

was more likely to judge harshly our subject?

会更有可能质疑他人呢？

Paradoxically, it was the group

自相矛盾的是，

who had posted similar information,

那是发布了类似信息的组，

an example of moral dissonance.

这就是一个与个人道德相悖的例子。

Now you may be thinking,

大家可能现在会想，

this does not apply to me,

这跟我无关，

because I have nothing to hide.

因为我没有什么可隐藏的。

But in fact, privacy is not about

但实际上，隐私不是说

having something negative to hide.

你有什么坏事情要隐藏。

Imagine that you are the H.R. director

想象一下你是某机构人事部的主管，

of a certain organization, and you receive résumés,

你收到一些简历，

and you decide to find more information about the candidates.

然后你决定寻找更多的关于这些应征者的信息。

Therefore, you Google their names

然后，你就在谷歌上搜他们的名字

and in a certain universe,

在某种情形下，

you find this information.

你找到这个信息。

Or in a parallel universe, you find this information.

或者在一个平行的空间里，你找到了这个信息。

Do you think that you would be equally likely

你认为你会公平的

to call either candidate for an interview?

给任何一个应征者面试的机会吗？

If you think so, then you are not

如果你是这样想的话，

like the U.S. employers who are, in fact,

那么你与美国的老板们不同,

part of our experiment, meaning we did exactly that.

实际上，我们就是用了这些老板做的这个试验。

We created Facebook profiles, manipulating traits,

我们建立了一些脸书帐号，编制了一些信息，

then we started sending out résumés to companies in the U.S.,

然后我们开始给他们发简历，

and we detected, we monitored,

此后我们监控着,

whether they were searching for our candidates,

到底这些公司会不会搜索我们的应征者，

and whether they were acting on the information

他们有没有对他们在社交网络上找到的信息

they found on social media. And they were.

有所举动。实际上他们确实这样做了。

Discrimination was happening through social media

对同等条件的应征者的歧视

for equally skilled candidates.

是正从社交网络收集的信息开始的。

Now marketers like us to believe

现在营销人员希望我们相信

that all information about us will always

关于我们的所有信息

be used in a manner which is in our favor.

永远都会以我们喜欢的方式被使用。

But think again. Why should that be always the case?

但是想想看,凭什么总会是这样？

In a movie which came out a few years ago,

在几年前出品的一部电影里，

"Minority Report," a famous scene

“少数派报告”，一个著名的镜头里

had Tom Cruise walk in a mall

是汤姆·克鲁斯在一个大厦里走着

and holographic personalized advertising

然后全息个性化的广告

would appear around him.

出现在他周围。

Now, that movie is set in 2054,

那部电影的背景年代是2054年，

about 40 years from now,

大约离现在还有40年，

and as exciting as that technology looks,

就像那个技术显示的一样让人兴奋，

it already vastly underestimates

它已经大大低估了

the amount of information that organizations

各种机构可以搜集到的

can gather about you, and how they can use it

关于你自己的信息，以及他们如何能利用这些信息

to influence you in a way that you will not even detect.

以一种你自己都无法预测到的方式来影响你。

So as an example, this is another experiment

举个例子，这是另一个

actually we are running, not yet completed.

我们正在做的未完成的试验。

Imagine that an organization has access

想象一下某个机构有

to your list of Facebook friends,

你的脸书朋友信息，

and through some kind of algorithm

通过某种算法

they can detect the two friends that you like the most.

他们可以找到两个你最喜欢的朋友。

And then they create, in real time,

然后，他们的即时创建出

a facial composite of these two friends.

这两个朋友的脸部信息结合体。

Now studies prior to ours have shown that people

在我们之前的研究显示

don't recognize any longer even themselves

人们在合成的脸部图片中

in facial composites, but they react

甚至不会识别出自己，

to those composites in a positive manner.

但是他们却对这些合成图片有好感。

So next time you are looking for a certain product,

那么下次你在浏览某个产品的时候，

and there is an ad suggesting you to buy it,

同时有个广告建议你买它，

it will not be just a standard spokesperson.

这就不会是一个标准的推销员,

It will be one of your friends,

却会变成你的朋友，

and you will not even know that this is happening.

而且你都不会意识到正在发生着什么。

Now the problem is that

现在的问题是

the current policy mechanisms we have

我们当下的保护个人信息

to protect ourselves from the abuses of personal information

不被滥用的政策法规

are like bringing a knife to a gunfight.

还十分薄弱。

One of these mechanisms is transparency,

其中的一个法规是透明性，

telling people what you are going to do with their data.

要告诉人们你将怎样使用这些数据。

And in principle, that's a very good thing.

理论上，这是非常好的事情。

It's necessary, but it is not sufficient.

这是必要的，但是却不完善。

Transparency can be misdirected.

透明性也会被误导。

You can tell people what you are going to do,

你会告诉人们你要做什么，

and then you still nudge them to disclose

然后你仍然试图诱导他们

arbitrary amounts of personal information.

给你任意数量的个人信息。

So in yet another experiment, this one with students,

那么在另一个实验里，这次我们让

we asked them to provide information

学生们给我们他们的

about their campus behavior,

学校表现信息

including pretty sensitive questions, such as this one.

这包括一些非常敏感的信息，比如这个。

[Have you ever cheated in an exam?]

Now to one group of subjects, we told them,

［你有没有在考试中做过弊？］

对其中一个组，我们告诉他们，

"Only other students will see your answers."

“只有其他的学生会看到你的答案。”

To another group of subjects, we told them,

而对另一组学生，我们说，

"Students and faculty will see your answers."

“学生和系里会看到你们的答案。”

Transparency. Notification. And sure enough, this worked,

透明度。预先声明。当然，这个奏效了，

in the sense that the first group of subjects

第一组学生

were much more likely to disclose than the second.

比第二组更愿意说出实情。

It makes sense, right?

很合理吧，不是吗？

But then we added the misdirection.

但是我们加了下面的误导。

We repeated the experiment with the same two groups,

我们在两组中重复做了这个实验，

this time adding a delay

这次我们

between the time we told subjects

在告诉他们我们如何

how we would use their data

使用这些数据

and the time we actually started answering the questions.

和让他们实际开始回答问题之间增加了一点延迟。

How long a delay do you think we had to add

大家认为这个延迟需要多久

in order to nullify the inhibitory effect

能让我们抵消掉之前的“系里也会看你们的答案”

of knowing that faculty would see your answers?

带来的抑制作用？

Ten minutes?

十分钟？

Five minutes?

五分钟？

One minute?

一分钟？

How about 15 seconds?

15秒怎么样？

Fifteen seconds were sufficient to have the two groups

只要15秒就会让两组

disclose the same amount of information,

提供同样数量的数据，

as if the second group now no longer cares

就好像第二组不再关心

for faculty reading their answers.

系里会不会看他们的答案一样。

Now I have to admit that this talk so far

到此为止，我得承认这个演讲

may sound exceedingly gloomy,

可能显得非常的郁闷,

but that is not my point.

但是这不是我的重点。

In fact, I want to share with you the fact that

实际上，我想分享的是我们还是有

there are alternatives.

其他办法的。

The way we are doing things now is not the only way

我们现在的处理方式不是唯一的，

they can done, and certainly not the best way

也绝对不是最好的。

they can be done.

也绝对不是最好的。

When someone tells you, "People don't care about privacy,"

当有人对你说，“大家不用关心隐私，”

consider whether the game has been designed

想想是不是因为事情已经被扭曲到

and rigged so that they cannot care about privacy,

他们不能再关心个人隐私了，

and coming to the realization that these manipulations occur

然后我们才意识到一切已被人操纵,

is already halfway through the process

已经逐渐侵入到

of being able to protect yourself.

自我保护的整个过程中。

When someone tells you that privacy is incompatible

当有人说隐私和大量信息带来的好处

with the benefits of big data,

无法兼得时，

consider that in the last 20 years,

想想过去的20年里，

researchers have created technologies

研究人员已经发明了

to allow virtually any electronic transactions

理论上使任何电子转帐

to take place in a more privacy-preserving manner.

更加安全保密的方式来进行的技术。

We can browse the Internet anonymously.

我们可以匿名的浏览网页。

We can send emails that can only be read

我们可以发送连美国国家安全局都不可以

by the intended recipient, not even the NSA.

读取的个人电子邮件，

We can have even privacy-preserving data mining.

我们甚至可以有保护隐私的数据挖掘。

In other words, we can have the benefits of big data

换句话说，我们可以在得到大量数据的同时

while protecting privacy.

仍能保护个人隐私。

Of course, these technologies imply a shifting

当然，这些技术的应用意味着

of cost and revenues

在数据拥有者们和

between data holders and data subjects,

数据对象们之间将有花费和收入的变化，

which is why, perhaps, you don't hear more about them.

也许这可能就是我们为什么没怎么听说过这些技术的原因。

Which brings me back to the Garden of Eden.

让我再回到伊甸园。

There is a second privacy interpretation

关于伊甸园的故事

of the story of the Garden of Eden

还有第二个关于隐私的解释

which doesn't have to do with the issue

这跟亚当和夏娃

of Adam and Eve feeling naked

的赤裸和羞耻

and feeling ashamed.

没有任何关系。

You can find echoes of this interpretation

大家可以在

in John Milton's "Paradise Lost."

约翰·弥尔顿的“失乐园”里看到类似的解释。

In the garden, Adam and Eve are materially content.

在伊甸园里，亚当和夏娃是物质上的满足。

They're happy. They are satisfied.

他们很开心，也很满足。

However, they also lack knowledge

但是，他们没有知识

and self-awareness.

和自觉性。

The moment they eat the aptly named

当他们吃到

fruit of knowledge,

智慧之果时，

that's when they discover themselves.

其实是他们发现自我的时刻。

They become aware. They achieve autonomy.

他们变得自觉，实现了自主。

The price to pay, however, is leaving the garden.

但是代价却是，离开伊甸园。

So privacy, in a way, is both the means

那么，隐私，换句话说，就是

and the price to pay for freedom.

为了得到自由必须付出的代价。

Again, marketers tell us

再次，营销人员告诉我们

that big data and social media

大量数据和社交网络

are not just a paradise of profit for them,

并不仅仅是为他们谋福利的天堂，

but a Garden of Eden for the rest of us.

同时也是我们所有人的伊甸园。

We get free content.

我们得到免费的信息。

We get to play Angry Birds. We get targeted apps.

我们可以玩愤怒的小鸟。我们得到适合自己的应用。

But in fact, in a few years, organizations

但实际上，在几年内，各种机构

will know so much about us,

就会因为知道这么多关于我们的信息，

they will be able to infer our desires

进而可以在我们知道自己想要做什么之前

before we even form them, and perhaps

就可以诱导我们的想法，或许

buy products on our behalf

在我们知道自己是不是真的需要某个商品之前

before we even know we need them.

就以我们自己的名义把它买下来了。

Now there was one English author

有一个英国作家

who anticipated this kind of future

预测到了这种未来

where we would trade away

就是我们会用自己的自主

our autonomy and freedom for comfort.

和自由来换来舒适安逸。

Even more so than George Orwell,

甚至超过了乔治·奥威尔，

the author is, of course, Aldous Huxley.

这个作家当然是赫胥黎。

In "Brave New World," he imagines a society

在“美丽新世界”里，他想象了一个社会:

where technologies that we created

人们发明了原本是为了得到

originally for freedom

自由的一种技术,

end up coercing us.

最终反被此技术所奴役。

However, in the book, he also offers us a way out

然而，在这本书里，他同样给我们指出了一条

of that society, similar to the path

突破这个社会的道路，

that Adam and Eve had to follow to leave the garden.

跟亚当和夏娃不得不离开伊甸园的道路类似。

In the words of the Savage,

用野人的话说，

regaining autonomy and freedom is possible,

重获自主和自由是可能的，

although the price to pay is steep.

尽管代价惨重。

So I do believe that one of the defining fights

因此我相信当今

of our times will be the fight

具有决定性的战役之一

for the control over personal information,

就是控制个人信息之战，

the fight over whether big data will become a force

决定大量数据是否会变成帮助获得自由

for freedom,

的武器，

rather than a force which will hiddenly manipulate us.

还是暗中操纵我们的工具。

Right now, many of us

现在，我们中的大多数

do not even know that the fight is going on,

甚至不知道战斗已经打响了，

but it is, whether you like it or not.

但这是真的，不管你喜欢不喜欢。

And at the risk of playing the serpent,

冒着打草惊蛇的危险，

I will tell you that the tools for the fight

我告诉大家战斗的武器就在这里，

are here, the awareness of what is going on,

那就是意识到正在发生着什么，

and in your hands,

就在你手中，

just a few clicks away.

只需几次点击。

Thank you.

谢谢大家。

(Applause)

（掌声）